/* will give out up to bits_wanted contiguous bits. */
static int ocfs2_claim_suballoc_bits(struct ocfs2_super *osb,
struct ocfs2_alloc_context *ac,
handle_t *handle,
u32 bits_wanted,
u32 min_bits,
u16 *bit_off,
unsigned int *num_bits,
u64 *bg_blkno)
{
int status;
u16 victim, i;
u16 bits_left = 0;
u64 hint_blkno = ac->ac_last_group;
struct ocfs2_chain_list *cl;
struct ocfs2_dinode *fe;
mlog_entry_void();
BUG_ON(ac->ac_bits_given >= ac->ac_bits_wanted);
BUG_ON(bits_wanted > (ac->ac_bits_wanted - ac->ac_bits_given));
BUG_ON(!ac->ac_bh);
fe = (struct ocfs2_dinode *) ac->ac_bh->b_data;
/* The bh was validated by the inode read during
* ocfs2_reserve_suballoc_bits(). Any corruption is a code bug. */
BUG_ON(!OCFS2_IS_VALID_DINODE(fe));
if (le32_to_cpu(fe->id1.bitmap1.i_used) >=
le32_to_cpu(fe->id1.bitmap1.i_total)) {
ocfs2_error(osb->sb, "Chain allocator dinode %llu has %u used "
"bits but only %u total."
static int ocfs2_journal_toggle_dirty(struct ocfs2_super *osb,
int dirty)
{
int status;
unsigned int flags;
struct ocfs2_journal *journal = osb->journal;
struct buffer_head *bh = journal->j_bh;
struct ocfs2_dinode *fe;
mlog_entry_void();
fe = (struct ocfs2_dinode *)bh->b_data;
if (!OCFS2_IS_VALID_DINODE(fe)) {
/* This is called from startup/shutdown which will
* handle the errors in a specific manner, so no need
* to call ocfs2_error() here. */
mlog(ML_ERROR, "Journal dinode %llu has invalid "
"signature: %.*s",
(unsigned long long)le64_to_cpu(fe->i_blkno), 7,
fe->i_signature);
status = -EIO;
goto out;
}
flags = le32_to_cpu(fe->id1.journal1.ij_flags);
if (dirty)
flags |= OCFS2_JOURNAL_DIRTY_FL;
else
flags &= ~OCFS2_JOURNAL_DIRTY_FL;
fe->id1.journal1.ij_flags = cpu_to_le32(flags);
status = ocfs2_write_block(osb, bh, journal->j_inode);
if (status < 0)
mlog_errno(status);
out:
mlog_exit(status);
return status;
}
int ocfs2_validate_inode_block(struct super_block *sb,
struct buffer_head *bh)
{
int rc;
struct ocfs2_dinode *di = (struct ocfs2_dinode *)bh->b_data;
trace_ocfs2_validate_inode_block((unsigned long long)bh->b_blocknr);
BUG_ON(!buffer_uptodate(bh));
/*
* If the ecc fails, we return the error but otherwise
* leave the filesystem running. We know any error is
* local to this block.
*/
rc = ocfs2_validate_meta_ecc(sb, bh->b_data, &di->i_check);
if (rc) {
mlog(ML_ERROR, "Checksum failed for dinode %llu\n",
(unsigned long long)bh->b_blocknr);
goto bail;
}
/*
* Errors after here are fatal.
*/
rc = -EINVAL;
if (!OCFS2_IS_VALID_DINODE(di)) {
ocfs2_error(sb, "Invalid dinode #%llu: signature = %.*s\n",
(unsigned long long)bh->b_blocknr, 7,
di->i_signature);
goto bail;
}
if (le64_to_cpu(di->i_blkno) != bh->b_blocknr) {
ocfs2_error(sb, "Invalid dinode #%llu: i_blkno is %llu\n",
(unsigned long long)bh->b_blocknr,
(unsigned long long)le64_to_cpu(di->i_blkno));
goto bail;
}
if (!(di->i_flags & cpu_to_le32(OCFS2_VALID_FL))) {
ocfs2_error(sb,
"Invalid dinode #%llu: OCFS2_VALID_FL not set\n",
(unsigned long long)bh->b_blocknr);
goto bail;
}
if (le32_to_cpu(di->i_fs_generation) !=
OCFS2_SB(sb)->fs_generation) {
ocfs2_error(sb,
"Invalid dinode #%llu: fs_generation is %u\n",
(unsigned long long)bh->b_blocknr,
le32_to_cpu(di->i_fs_generation));
goto bail;
}
rc = 0;
bail:
return rc;
}
/*
* This lookup actually will read from disk. It has one invariant:
* It will never re-traverse blocks. This means that all inserts should
* be new regions or more granular regions (both allowed by insert).
*/
static int ocfs2_extent_map_lookup_read(struct inode *inode,
u32 cpos,
u32 clusters,
struct ocfs2_extent_map_entry **ret_ent)
{
int ret;
u64 blkno;
struct ocfs2_extent_map *em = &OCFS2_I(inode)->ip_map;
struct ocfs2_extent_map_entry *ent;
struct buffer_head *bh = NULL;
struct ocfs2_extent_block *eb;
struct ocfs2_dinode *di;
struct ocfs2_extent_list *el;
spin_lock(&OCFS2_I(inode)->ip_lock);
ent = ocfs2_extent_map_lookup(em, cpos, clusters, NULL, NULL);
if (ent) {
if (!ent->e_tree_depth) {
spin_unlock(&OCFS2_I(inode)->ip_lock);
*ret_ent = ent;
return 0;
}
blkno = le64_to_cpu(ent->e_rec.e_blkno);
spin_unlock(&OCFS2_I(inode)->ip_lock);
ret = ocfs2_read_block(OCFS2_SB(inode->i_sb), blkno, &bh,
OCFS2_BH_CACHED, inode);
if (ret) {
mlog_errno(ret);
if (bh)
brelse(bh);
return ret;
}
eb = (struct ocfs2_extent_block *)bh->b_data;
if (!OCFS2_IS_VALID_EXTENT_BLOCK(eb)) {
OCFS2_RO_ON_INVALID_EXTENT_BLOCK(inode->i_sb, eb);
brelse(bh);
return -EIO;
}
el = &eb->h_list;
} else {
spin_unlock(&OCFS2_I(inode)->ip_lock);
ret = ocfs2_read_block(OCFS2_SB(inode->i_sb),
OCFS2_I(inode)->ip_blkno, &bh,
OCFS2_BH_CACHED, inode);
if (ret) {
mlog_errno(ret);
if (bh)
brelse(bh);
return ret;
}
di = (struct ocfs2_dinode *)bh->b_data;
if (!OCFS2_IS_VALID_DINODE(di)) {
brelse(bh);
OCFS2_RO_ON_INVALID_DINODE(inode->i_sb, di);
return -EIO;
}
el = &di->id2.i_list;
}
ret = ocfs2_extent_map_find_leaf(inode, cpos, clusters, el);
brelse(bh);
if (ret) {
mlog_errno(ret);
return ret;
}
ent = ocfs2_extent_map_lookup(em, cpos, clusters, NULL, NULL);
if (!ent) {
ret = -ESRCH;
mlog_errno(ret);
return ret;
}
/* FIXME: Make sure this isn't a corruption */
BUG_ON(ent->e_tree_depth);
*ret_ent = ent;
return 0;
}
static int ocfs2_symlink_get_block(struct inode *inode, sector_t iblock,
struct buffer_head *bh_result, int create)
{
int err = -EIO;
int status;
struct ocfs2_dinode *fe = NULL;
struct buffer_head *bh = NULL;
struct buffer_head *buffer_cache_bh = NULL;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
void *kaddr;
mlog_entry("(0x%p, %llu, 0x%p, %d)\n", inode,
(unsigned long long)iblock, bh_result, create);
BUG_ON(ocfs2_inode_is_fast_symlink(inode));
if ((iblock << inode->i_sb->s_blocksize_bits) > PATH_MAX + 1) {
mlog(ML_ERROR, "block offset > PATH_MAX: %llu",
(unsigned long long)iblock);
goto bail;
}
status = ocfs2_read_block(OCFS2_SB(inode->i_sb),
OCFS2_I(inode)->ip_blkno,
&bh, OCFS2_BH_CACHED, inode);
if (status < 0) {
mlog_errno(status);
goto bail;
}
fe = (struct ocfs2_dinode *) bh->b_data;
if (!OCFS2_IS_VALID_DINODE(fe)) {
mlog(ML_ERROR, "Invalid dinode #%llu: signature = %.*s\n",
(unsigned long long)le64_to_cpu(fe->i_blkno), 7,
fe->i_signature);
goto bail;
}
if ((u64)iblock >= ocfs2_clusters_to_blocks(inode->i_sb,
le32_to_cpu(fe->i_clusters))) {
mlog(ML_ERROR, "block offset is outside the allocated size: "
"%llu\n", (unsigned long long)iblock);
goto bail;
}
/* We don't use the page cache to create symlink data, so if
* need be, copy it over from the buffer cache. */
if (!buffer_uptodate(bh_result) && ocfs2_inode_is_new(inode)) {
u64 blkno = le64_to_cpu(fe->id2.i_list.l_recs[0].e_blkno) +
iblock;
buffer_cache_bh = sb_getblk(osb->sb, blkno);
if (!buffer_cache_bh) {
mlog(ML_ERROR, "couldn't getblock for symlink!\n");
goto bail;
}
/* we haven't locked out transactions, so a commit
* could've happened. Since we've got a reference on
* the bh, even if it commits while we're doing the
* copy, the data is still good. */
if (buffer_jbd(buffer_cache_bh)
&& ocfs2_inode_is_new(inode)) {
kaddr = kmap_atomic(bh_result->b_page, KM_USER0);
if (!kaddr) {
mlog(ML_ERROR, "couldn't kmap!\n");
goto bail;
}
memcpy(kaddr + (bh_result->b_size * iblock),
buffer_cache_bh->b_data,
bh_result->b_size);
kunmap_atomic(kaddr, KM_USER0);
set_buffer_uptodate(bh_result);
}
brelse(buffer_cache_bh);
}
map_bh(bh_result, inode->i_sb,
le64_to_cpu(fe->id2.i_list.l_recs[0].e_blkno) + iblock);
err = 0;
bail:
if (bh)
brelse(bh);
mlog_exit(err);
return err;
}
static int ocfs2_read_locked_inode(struct inode *inode,
struct ocfs2_find_inode_args *args)
{
struct super_block *sb;
struct ocfs2_super *osb;
struct ocfs2_dinode *fe;
struct buffer_head *bh = NULL;
int status, can_lock;
u32 generation = 0;
mlog_entry("(0x%p, 0x%p)\n", inode, args);
status = -EINVAL;
if (inode == NULL || inode->i_sb == NULL) {
mlog(ML_ERROR, "bad inode\n");
return status;
}
sb = inode->i_sb;
osb = OCFS2_SB(sb);
if (!args) {
mlog(ML_ERROR, "bad inode args\n");
make_bad_inode(inode);
return status;
}
/*
* To improve performance of cold-cache inode stats, we take
* the cluster lock here if possible.
*
* Generally, OCFS2 never trusts the contents of an inode
* unless it's holding a cluster lock, so taking it here isn't
* a correctness issue as much as it is a performance
* improvement.
*
* There are three times when taking the lock is not a good idea:
*
* 1) During startup, before we have initialized the DLM.
*
* 2) If we are reading certain system files which never get
* cluster locks (local alloc, truncate log).
*
* 3) If the process doing the iget() is responsible for
* orphan dir recovery. We're holding the orphan dir lock and
* can get into a deadlock with another process on another
* node in ->delete_inode().
*
* #1 and #2 can be simply solved by never taking the lock
* here for system files (which are the only type we read
* during mount). It's a heavier approach, but our main
* concern is user-accesible files anyway.
*
* #3 works itself out because we'll eventually take the
* cluster lock before trusting anything anyway.
*/
can_lock = !(args->fi_flags & OCFS2_FI_FLAG_SYSFILE)
&& !(args->fi_flags & OCFS2_FI_FLAG_ORPHAN_RECOVERY)
&& !ocfs2_mount_local(osb);
/*
* To maintain backwards compatibility with older versions of
* ocfs2-tools, we still store the generation value for system
* files. The only ones that actually matter to userspace are
* the journals, but it's easier and inexpensive to just flag
* all system files similarly.
*/
if (args->fi_flags & OCFS2_FI_FLAG_SYSFILE)
generation = osb->fs_generation;
ocfs2_inode_lock_res_init(&OCFS2_I(inode)->ip_meta_lockres,
OCFS2_LOCK_TYPE_META,
generation, inode);
ocfs2_inode_lock_res_init(&OCFS2_I(inode)->ip_open_lockres,
OCFS2_LOCK_TYPE_OPEN,
0, inode);
if (can_lock) {
status = ocfs2_open_lock(inode);
if (status) {
make_bad_inode(inode);
mlog_errno(status);
return status;
}
status = ocfs2_meta_lock(inode, NULL, 0);
if (status) {
make_bad_inode(inode);
mlog_errno(status);
return status;
}
}
if (args->fi_flags & OCFS2_FI_FLAG_ORPHAN_RECOVERY) {
status = ocfs2_try_open_lock(inode, 0);
if (status) {
make_bad_inode(inode);
return status;
}
}
status = ocfs2_read_block(osb, args->fi_blkno, &bh, 0,
can_lock ? inode : NULL);
if (status < 0) {
mlog_errno(status);
goto bail;
}
status = -EINVAL;
fe = (struct ocfs2_dinode *) bh->b_data;
if (!OCFS2_IS_VALID_DINODE(fe)) {
mlog(0, "Invalid dinode #%llu: signature = %.*s\n",
(unsigned long long)args->fi_blkno, 7,
fe->i_signature);
goto bail;
}
/*
* This is a code bug. Right now the caller needs to
* understand whether it is asking for a system file inode or
* not so the proper lock names can be built.
*/
mlog_bug_on_msg(!!(fe->i_flags & cpu_to_le32(OCFS2_SYSTEM_FL)) !=
!!(args->fi_flags & OCFS2_FI_FLAG_SYSFILE),
"Inode %llu: system file state is ambigous\n",
(unsigned long long)args->fi_blkno);
if (S_ISCHR(le16_to_cpu(fe->i_mode)) ||
S_ISBLK(le16_to_cpu(fe->i_mode)))
inode->i_rdev = huge_decode_dev(le64_to_cpu(fe->id1.dev1.i_rdev));
if (ocfs2_populate_inode(inode, fe, 0) < 0)
goto bail;
BUG_ON(args->fi_blkno != le64_to_cpu(fe->i_blkno));
status = 0;
bail:
if (can_lock)
ocfs2_meta_unlock(inode, 0);
if (status < 0)
make_bad_inode(inode);
if (args && bh)
brelse(bh);
mlog_exit(status);
return status;
}
int ocfs2_populate_inode(struct inode *inode, struct ocfs2_dinode *fe,
int create_ino)
{
struct super_block *sb;
struct ocfs2_super *osb;
int status = -EINVAL;
mlog_entry("(0x%p, size:%llu)\n", inode,
(unsigned long long)le64_to_cpu(fe->i_size));
sb = inode->i_sb;
osb = OCFS2_SB(sb);
/* this means that read_inode cannot create a superblock inode
* today. change if needed. */
if (!OCFS2_IS_VALID_DINODE(fe) ||
!(fe->i_flags & cpu_to_le32(OCFS2_VALID_FL))) {
mlog(0, "Invalid dinode: i_ino=%lu, i_blkno=%llu, "
"signature = %.*s, flags = 0x%x\n",
inode->i_ino,
(unsigned long long)le64_to_cpu(fe->i_blkno), 7,
fe->i_signature, le32_to_cpu(fe->i_flags));
goto bail;
}
if (le32_to_cpu(fe->i_fs_generation) != osb->fs_generation) {
mlog(ML_ERROR, "file entry generation does not match "
"superblock! osb->fs_generation=%x, "
"fe->i_fs_generation=%x\n",
osb->fs_generation, le32_to_cpu(fe->i_fs_generation));
goto bail;
}
OCFS2_I(inode)->ip_clusters = le32_to_cpu(fe->i_clusters);
OCFS2_I(inode)->ip_attr = le32_to_cpu(fe->i_attr);
OCFS2_I(inode)->ip_dyn_features = le16_to_cpu(fe->i_dyn_features);
inode->i_version = 1;
inode->i_generation = le32_to_cpu(fe->i_generation);
inode->i_rdev = huge_decode_dev(le64_to_cpu(fe->id1.dev1.i_rdev));
inode->i_mode = le16_to_cpu(fe->i_mode);
inode->i_uid = le32_to_cpu(fe->i_uid);
inode->i_gid = le32_to_cpu(fe->i_gid);
/* Fast symlinks will have i_size but no allocated clusters. */
if (S_ISLNK(inode->i_mode) && !fe->i_clusters)
inode->i_blocks = 0;
else
inode->i_blocks = ocfs2_inode_sector_count(inode);
inode->i_mapping->a_ops = &ocfs2_aops;
inode->i_atime.tv_sec = le64_to_cpu(fe->i_atime);
inode->i_atime.tv_nsec = le32_to_cpu(fe->i_atime_nsec);
inode->i_mtime.tv_sec = le64_to_cpu(fe->i_mtime);
inode->i_mtime.tv_nsec = le32_to_cpu(fe->i_mtime_nsec);
inode->i_ctime.tv_sec = le64_to_cpu(fe->i_ctime);
inode->i_ctime.tv_nsec = le32_to_cpu(fe->i_ctime_nsec);
if (OCFS2_I(inode)->ip_blkno != le64_to_cpu(fe->i_blkno))
mlog(ML_ERROR,
"ip_blkno %llu != i_blkno %llu!\n",
(unsigned long long)OCFS2_I(inode)->ip_blkno,
(unsigned long long)le64_to_cpu(fe->i_blkno));
inode->i_nlink = le16_to_cpu(fe->i_links_count);
if (fe->i_flags & cpu_to_le32(OCFS2_SYSTEM_FL))
OCFS2_I(inode)->ip_flags |= OCFS2_INODE_SYSTEM_FILE;
if (fe->i_flags & cpu_to_le32(OCFS2_LOCAL_ALLOC_FL)) {
OCFS2_I(inode)->ip_flags |= OCFS2_INODE_BITMAP;
mlog(0, "local alloc inode: i_ino=%lu\n", inode->i_ino);
} else if (fe->i_flags & cpu_to_le32(OCFS2_BITMAP_FL)) {
OCFS2_I(inode)->ip_flags |= OCFS2_INODE_BITMAP;
} else if (fe->i_flags & cpu_to_le32(OCFS2_SUPER_BLOCK_FL)) {
mlog(0, "superblock inode: i_ino=%lu\n", inode->i_ino);
/* we can't actually hit this as read_inode can't
* handle superblocks today ;-) */
BUG();
}
switch (inode->i_mode & S_IFMT) {
case S_IFREG:
inode->i_fop = &ocfs2_fops;
inode->i_op = &ocfs2_file_iops;
i_size_write(inode, le64_to_cpu(fe->i_size));
break;
case S_IFDIR:
inode->i_op = &ocfs2_dir_iops;
inode->i_fop = &ocfs2_dops;
i_size_write(inode, le64_to_cpu(fe->i_size));
break;
case S_IFLNK:
if (ocfs2_inode_is_fast_symlink(inode))
inode->i_op = &ocfs2_fast_symlink_inode_operations;
else
inode->i_op = &ocfs2_symlink_inode_operations;
i_size_write(inode, le64_to_cpu(fe->i_size));
break;
default:
inode->i_op = &ocfs2_special_file_iops;
init_special_inode(inode, inode->i_mode,
inode->i_rdev);
break;
}
if (create_ino) {
inode->i_ino = ino_from_blkno(inode->i_sb,
le64_to_cpu(fe->i_blkno));
/*
* If we ever want to create system files from kernel,
* the generation argument to
* ocfs2_inode_lock_res_init() will have to change.
*/
BUG_ON(le32_to_cpu(fe->i_flags) & OCFS2_SYSTEM_FL);
ocfs2_inode_lock_res_init(&OCFS2_I(inode)->ip_meta_lockres,
OCFS2_LOCK_TYPE_META, 0, inode);
ocfs2_inode_lock_res_init(&OCFS2_I(inode)->ip_open_lockres,
OCFS2_LOCK_TYPE_OPEN, 0, inode);
}
ocfs2_inode_lock_res_init(&OCFS2_I(inode)->ip_rw_lockres,
OCFS2_LOCK_TYPE_RW, inode->i_generation,
inode);
ocfs2_inode_lock_res_init(&OCFS2_I(inode)->ip_data_lockres,
OCFS2_LOCK_TYPE_DATA, inode->i_generation,
inode);
ocfs2_set_inode_flags(inode);
status = 0;
bail:
mlog_exit(status);
return status;
}
static int ocfs2_extend_allocation(struct inode *inode,
u32 clusters_to_add)
{
int status = 0;
int restart_func = 0;
int drop_alloc_sem = 0;
int credits;
u32 prev_clusters, logical_start;
struct buffer_head *bh = NULL;
struct ocfs2_dinode *fe = NULL;
handle_t *handle = NULL;
struct ocfs2_alloc_context *data_ac = NULL;
struct ocfs2_alloc_context *meta_ac = NULL;
enum ocfs2_alloc_restarted why;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
mlog_entry("(clusters_to_add = %u)\n", clusters_to_add);
/*
* This function only exists for file systems which don't
* support holes.
*/
BUG_ON(ocfs2_sparse_alloc(osb));
status = ocfs2_read_block(osb, OCFS2_I(inode)->ip_blkno, &bh,
OCFS2_BH_CACHED, inode);
if (status < 0) {
mlog_errno(status);
goto leave;
}
fe = (struct ocfs2_dinode *) bh->b_data;
if (!OCFS2_IS_VALID_DINODE(fe)) {
OCFS2_RO_ON_INVALID_DINODE(inode->i_sb, fe);
status = -EIO;
goto leave;
}
logical_start = OCFS2_I(inode)->ip_clusters;
restart_all:
BUG_ON(le32_to_cpu(fe->i_clusters) != OCFS2_I(inode)->ip_clusters);
/* blocks peope in read/write from reading our allocation
* until we're done changing it. We depend on i_mutex to block
* other extend/truncate calls while we're here. Ordering wrt
* start_trans is important here -- always do it before! */
down_write(&OCFS2_I(inode)->ip_alloc_sem);
drop_alloc_sem = 1;
status = ocfs2_lock_allocators(inode, fe, clusters_to_add, &data_ac,
&meta_ac);
if (status) {
mlog_errno(status);
goto leave;
}
credits = ocfs2_calc_extend_credits(osb->sb, fe, clusters_to_add);
handle = ocfs2_start_trans(osb, credits);
if (IS_ERR(handle)) {
status = PTR_ERR(handle);
handle = NULL;
mlog_errno(status);
goto leave;
}
restarted_transaction:
/* reserve a write to the file entry early on - that we if we
* run out of credits in the allocation path, we can still
* update i_size. */
status = ocfs2_journal_access(handle, inode, bh,
OCFS2_JOURNAL_ACCESS_WRITE);
if (status < 0) {
mlog_errno(status);
goto leave;
}
prev_clusters = OCFS2_I(inode)->ip_clusters;
status = ocfs2_do_extend_allocation(osb,
inode,
&logical_start,
clusters_to_add,
bh,
handle,
data_ac,
meta_ac,
&why);
if ((status < 0) && (status != -EAGAIN)) {
if (status != -ENOSPC)
mlog_errno(status);
goto leave;
}
status = ocfs2_journal_dirty(handle, bh);
if (status < 0) {
mlog_errno(status);
goto leave;
}
spin_lock(&OCFS2_I(inode)->ip_lock);
clusters_to_add -= (OCFS2_I(inode)->ip_clusters - prev_clusters);
spin_unlock(&OCFS2_I(inode)->ip_lock);
if (why != RESTART_NONE && clusters_to_add) {
if (why == RESTART_META) {
mlog(0, "restarting function.\n");
restart_func = 1;
} else {
BUG_ON(why != RESTART_TRANS);
mlog(0, "restarting transaction.\n");
/* TODO: This can be more intelligent. */
credits = ocfs2_calc_extend_credits(osb->sb,
fe,
clusters_to_add);
status = ocfs2_extend_trans(handle, credits);
if (status < 0) {
/* handle still has to be committed at
* this point. */
status = -ENOMEM;
mlog_errno(status);
goto leave;
}
goto restarted_transaction;
}
}
mlog(0, "fe: i_clusters = %u, i_size=%llu\n",
le32_to_cpu(fe->i_clusters),
(unsigned long long)le64_to_cpu(fe->i_size));
mlog(0, "inode: ip_clusters=%u, i_size=%lld\n",
OCFS2_I(inode)->ip_clusters, i_size_read(inode));
leave:
if (drop_alloc_sem) {
up_write(&OCFS2_I(inode)->ip_alloc_sem);
drop_alloc_sem = 0;
}
if (handle) {
ocfs2_commit_trans(osb, handle);
handle = NULL;
}
if (data_ac) {
ocfs2_free_alloc_context(data_ac);
data_ac = NULL;
}
if (meta_ac) {
ocfs2_free_alloc_context(meta_ac);
meta_ac = NULL;
}
if ((!status) && restart_func) {
restart_func = 0;
goto restart_all;
}
if (bh) {
brelse(bh);
bh = NULL;
}
mlog_exit(status);
return status;
}
static int ocfs2_reserve_suballoc_bits(struct ocfs2_super *osb,
struct ocfs2_alloc_context *ac,
int type,
u32 slot,
u64 *last_alloc_group,
int flags)
{
int status;
u32 bits_wanted = ac->ac_bits_wanted;
struct inode *alloc_inode;
struct buffer_head *bh = NULL;
struct ocfs2_dinode *fe;
u32 free_bits;
alloc_inode = ocfs2_get_system_file_inode(osb, type, slot);
if (!alloc_inode) {
mlog_errno(-EINVAL);
return -EINVAL;
}
mutex_lock(&alloc_inode->i_mutex);
status = ocfs2_inode_lock(alloc_inode, &bh, 1);
if (status < 0) {
mutex_unlock(&alloc_inode->i_mutex);
iput(alloc_inode);
mlog_errno(status);
return status;
}
ac->ac_inode = alloc_inode;
ac->ac_alloc_slot = slot;
fe = (struct ocfs2_dinode *) bh->b_data;
/* The bh was validated by the inode read inside
* ocfs2_inode_lock(). Any corruption is a code bug. */
BUG_ON(!OCFS2_IS_VALID_DINODE(fe));
if (!(fe->i_flags & cpu_to_le32(OCFS2_CHAIN_FL))) {
status = ocfs2_error(alloc_inode->i_sb,
"Invalid chain allocator %llu\n",
(unsigned long long)le64_to_cpu(fe->i_blkno));
goto bail;
}
free_bits = le32_to_cpu(fe->id1.bitmap1.i_total) -
le32_to_cpu(fe->id1.bitmap1.i_used);
if (bits_wanted > free_bits) {
/* cluster bitmap never grows */
if (ocfs2_is_cluster_bitmap(alloc_inode)) {
trace_ocfs2_reserve_suballoc_bits_nospc(bits_wanted,
free_bits);
status = -ENOSPC;
goto bail;
}
if (!(flags & ALLOC_NEW_GROUP)) {
trace_ocfs2_reserve_suballoc_bits_no_new_group(
slot, bits_wanted, free_bits);
status = -ENOSPC;
goto bail;
}
status = ocfs2_block_group_alloc(osb, alloc_inode, bh,
ac->ac_max_block,
last_alloc_group, flags);
if (status < 0) {
if (status != -ENOSPC)
mlog_errno(status);
goto bail;
}
atomic_inc(&osb->alloc_stats.bg_extends);
/* You should never ask for this much metadata */
BUG_ON(bits_wanted >
(le32_to_cpu(fe->id1.bitmap1.i_total)
- le32_to_cpu(fe->id1.bitmap1.i_used)));
}
get_bh(bh);
ac->ac_bh = bh;
bail:
brelse(bh);
if (status)
mlog_errno(status);
return status;
}
static int ocfs2_filecheck_repair_inode_block(struct super_block *sb,
struct buffer_head *bh)
{
int changed = 0;
struct ocfs2_dinode *di = (struct ocfs2_dinode *)bh->b_data;
if (!ocfs2_filecheck_validate_inode_block(sb, bh))
return 0;
trace_ocfs2_filecheck_repair_inode_block(
(unsigned long long)bh->b_blocknr);
if (ocfs2_is_hard_readonly(OCFS2_SB(sb)) ||
ocfs2_is_soft_readonly(OCFS2_SB(sb))) {
mlog(ML_ERROR,
"Filecheck: cannot repair dinode #%llu "
"on readonly filesystem\n",
(unsigned long long)bh->b_blocknr);
return -OCFS2_FILECHECK_ERR_READONLY;
}
if (buffer_jbd(bh)) {
mlog(ML_ERROR,
"Filecheck: cannot repair dinode #%llu, "
"its buffer is in jbd\n",
(unsigned long long)bh->b_blocknr);
return -OCFS2_FILECHECK_ERR_INJBD;
}
if (!OCFS2_IS_VALID_DINODE(di)) {
/* Cannot fix invalid inode block */
return -OCFS2_FILECHECK_ERR_INVALIDINO;
}
if (!(di->i_flags & cpu_to_le32(OCFS2_VALID_FL))) {
/* Cannot just add VALID_FL flag back as a fix,
* need more things to check here.
*/
return -OCFS2_FILECHECK_ERR_VALIDFLAG;
}
if (le64_to_cpu(di->i_blkno) != bh->b_blocknr) {
di->i_blkno = cpu_to_le64(bh->b_blocknr);
changed = 1;
mlog(ML_ERROR,
"Filecheck: reset dinode #%llu: i_blkno to %llu\n",
(unsigned long long)bh->b_blocknr,
(unsigned long long)le64_to_cpu(di->i_blkno));
}
if (le32_to_cpu(di->i_fs_generation) !=
OCFS2_SB(sb)->fs_generation) {
di->i_fs_generation = cpu_to_le32(OCFS2_SB(sb)->fs_generation);
changed = 1;
mlog(ML_ERROR,
"Filecheck: reset dinode #%llu: fs_generation to %u\n",
(unsigned long long)bh->b_blocknr,
le32_to_cpu(di->i_fs_generation));
}
if (changed || ocfs2_validate_meta_ecc(sb, bh->b_data, &di->i_check)) {
ocfs2_compute_meta_ecc(sb, bh->b_data, &di->i_check);
mark_buffer_dirty(bh);
mlog(ML_ERROR,
"Filecheck: reset dinode #%llu: compute meta ecc\n",
(unsigned long long)bh->b_blocknr);
}
return 0;
}
static int ocfs2_filecheck_validate_inode_block(struct super_block *sb,
struct buffer_head *bh)
{
int rc = 0;
struct ocfs2_dinode *di = (struct ocfs2_dinode *)bh->b_data;
trace_ocfs2_filecheck_validate_inode_block(
(unsigned long long)bh->b_blocknr);
BUG_ON(!buffer_uptodate(bh));
/*
* Call ocfs2_validate_meta_ecc() first since it has ecc repair
* function, but we should not return error immediately when ecc
* validation fails, because the reason is quite likely the invalid
* inode number inputed.
*/
rc = ocfs2_validate_meta_ecc(sb, bh->b_data, &di->i_check);
if (rc) {
mlog(ML_ERROR,
"Filecheck: checksum failed for dinode %llu\n",
(unsigned long long)bh->b_blocknr);
rc = -OCFS2_FILECHECK_ERR_BLOCKECC;
}
if (!OCFS2_IS_VALID_DINODE(di)) {
mlog(ML_ERROR,
"Filecheck: invalid dinode #%llu: signature = %.*s\n",
(unsigned long long)bh->b_blocknr, 7, di->i_signature);
rc = -OCFS2_FILECHECK_ERR_INVALIDINO;
goto bail;
} else if (rc)
goto bail;
if (le64_to_cpu(di->i_blkno) != bh->b_blocknr) {
mlog(ML_ERROR,
"Filecheck: invalid dinode #%llu: i_blkno is %llu\n",
(unsigned long long)bh->b_blocknr,
(unsigned long long)le64_to_cpu(di->i_blkno));
rc = -OCFS2_FILECHECK_ERR_BLOCKNO;
goto bail;
}
if (!(di->i_flags & cpu_to_le32(OCFS2_VALID_FL))) {
mlog(ML_ERROR,
"Filecheck: invalid dinode #%llu: OCFS2_VALID_FL "
"not set\n",
(unsigned long long)bh->b_blocknr);
rc = -OCFS2_FILECHECK_ERR_VALIDFLAG;
goto bail;
}
if (le32_to_cpu(di->i_fs_generation) !=
OCFS2_SB(sb)->fs_generation) {
mlog(ML_ERROR,
"Filecheck: invalid dinode #%llu: fs_generation is %u\n",
(unsigned long long)bh->b_blocknr,
le32_to_cpu(di->i_fs_generation));
rc = -OCFS2_FILECHECK_ERR_GENERATION;
}
bail:
return rc;
}
static int ocfs2_relink_block_group(handle_t *handle,
struct inode *alloc_inode,
struct buffer_head *fe_bh,
struct buffer_head *bg_bh,
struct buffer_head *prev_bg_bh,
u16 chain)
{
int status;
/* there is a really tiny chance the journal calls could fail,
* but we wouldn't want inconsistent blocks in *any* case. */
u64 fe_ptr, bg_ptr, prev_bg_ptr;
struct ocfs2_dinode *fe = (struct ocfs2_dinode *) fe_bh->b_data;
struct ocfs2_group_desc *bg = (struct ocfs2_group_desc *) bg_bh->b_data;
struct ocfs2_group_desc *prev_bg = (struct ocfs2_group_desc *) prev_bg_bh->b_data;
if (!OCFS2_IS_VALID_DINODE(fe)) {
OCFS2_RO_ON_INVALID_DINODE(alloc_inode->i_sb, fe);
status = -EIO;
goto out;
}
if (!OCFS2_IS_VALID_GROUP_DESC(bg)) {
OCFS2_RO_ON_INVALID_GROUP_DESC(alloc_inode->i_sb, bg);
status = -EIO;
goto out;
}
if (!OCFS2_IS_VALID_GROUP_DESC(prev_bg)) {
OCFS2_RO_ON_INVALID_GROUP_DESC(alloc_inode->i_sb, prev_bg);
status = -EIO;
goto out;
}
mlog(0, "Suballoc %llu, chain %u, move group %llu to top, prev = %llu\n",
(unsigned long long)le64_to_cpu(fe->i_blkno), chain,
(unsigned long long)le64_to_cpu(bg->bg_blkno),
(unsigned long long)le64_to_cpu(prev_bg->bg_blkno));
fe_ptr = le64_to_cpu(fe->id2.i_chain.cl_recs[chain].c_blkno);
bg_ptr = le64_to_cpu(bg->bg_next_group);
prev_bg_ptr = le64_to_cpu(prev_bg->bg_next_group);
status = ocfs2_journal_access(handle, alloc_inode, prev_bg_bh,
OCFS2_JOURNAL_ACCESS_WRITE);
if (status < 0) {
mlog_errno(status);
goto out_rollback;
}
prev_bg->bg_next_group = bg->bg_next_group;
status = ocfs2_journal_dirty(handle, prev_bg_bh);
if (status < 0) {
mlog_errno(status);
goto out_rollback;
}
status = ocfs2_journal_access(handle, alloc_inode, bg_bh,
OCFS2_JOURNAL_ACCESS_WRITE);
if (status < 0) {
mlog_errno(status);
goto out_rollback;
}
bg->bg_next_group = fe->id2.i_chain.cl_recs[chain].c_blkno;
status = ocfs2_journal_dirty(handle, bg_bh);
if (status < 0) {
mlog_errno(status);
goto out_rollback;
}
status = ocfs2_journal_access(handle, alloc_inode, fe_bh,
OCFS2_JOURNAL_ACCESS_WRITE);
if (status < 0) {
mlog_errno(status);
goto out_rollback;
}
fe->id2.i_chain.cl_recs[chain].c_blkno = bg->bg_blkno;
status = ocfs2_journal_dirty(handle, fe_bh);
if (status < 0) {
mlog_errno(status);
goto out_rollback;
}
status = 0;
out_rollback:
if (status < 0) {
fe->id2.i_chain.cl_recs[chain].c_blkno = cpu_to_le64(fe_ptr);
bg->bg_next_group = cpu_to_le64(bg_ptr);
prev_bg->bg_next_group = cpu_to_le64(prev_bg_ptr);
}
out:
mlog_exit(status);
return status;
}
static int ocfs2_reserve_suballoc_bits(struct ocfs2_super *osb,
struct ocfs2_alloc_context *ac,
int type,
u32 slot)
{
int status;
u32 bits_wanted = ac->ac_bits_wanted;
struct inode *alloc_inode;
struct buffer_head *bh = NULL;
struct ocfs2_dinode *fe;
u32 free_bits;
mlog_entry_void();
alloc_inode = ocfs2_get_system_file_inode(osb, type, slot);
if (!alloc_inode) {
mlog_errno(-EINVAL);
return -EINVAL;
}
mutex_lock(&alloc_inode->i_mutex);
status = ocfs2_meta_lock(alloc_inode, &bh, 1);
if (status < 0) {
mutex_unlock(&alloc_inode->i_mutex);
iput(alloc_inode);
mlog_errno(status);
return status;
}
ac->ac_inode = alloc_inode;
fe = (struct ocfs2_dinode *) bh->b_data;
if (!OCFS2_IS_VALID_DINODE(fe)) {
OCFS2_RO_ON_INVALID_DINODE(alloc_inode->i_sb, fe);
status = -EIO;
goto bail;
}
if (!(fe->i_flags & cpu_to_le32(OCFS2_CHAIN_FL))) {
ocfs2_error(alloc_inode->i_sb, "Invalid chain allocator %llu",
(unsigned long long)le64_to_cpu(fe->i_blkno));
status = -EIO;
goto bail;
}
free_bits = le32_to_cpu(fe->id1.bitmap1.i_total) -
le32_to_cpu(fe->id1.bitmap1.i_used);
if (bits_wanted > free_bits) {
/* cluster bitmap never grows */
if (ocfs2_is_cluster_bitmap(alloc_inode)) {
mlog(0, "Disk Full: wanted=%u, free_bits=%u\n",
bits_wanted, free_bits);
status = -ENOSPC;
goto bail;
}
status = ocfs2_block_group_alloc(osb, alloc_inode, bh);
if (status < 0) {
if (status != -ENOSPC)
mlog_errno(status);
goto bail;
}
atomic_inc(&osb->alloc_stats.bg_extends);
/* You should never ask for this much metadata */
BUG_ON(bits_wanted >
(le32_to_cpu(fe->id1.bitmap1.i_total)
- le32_to_cpu(fe->id1.bitmap1.i_used)));
}
get_bh(bh);
ac->ac_bh = bh;
bail:
if (bh)
brelse(bh);
mlog_exit(status);
return status;
}
/*
* expects the suballoc inode to already be locked.
*/
static int ocfs2_free_suballoc_bits(handle_t *handle,
struct inode *alloc_inode,
struct buffer_head *alloc_bh,
unsigned int start_bit,
u64 bg_blkno,
unsigned int count)
{
int status = 0;
u32 tmp_used;
struct ocfs2_super *osb = OCFS2_SB(alloc_inode->i_sb);
struct ocfs2_dinode *fe = (struct ocfs2_dinode *) alloc_bh->b_data;
struct ocfs2_chain_list *cl = &fe->id2.i_chain;
struct buffer_head *group_bh = NULL;
struct ocfs2_group_desc *group;
mlog_entry_void();
if (!OCFS2_IS_VALID_DINODE(fe)) {
OCFS2_RO_ON_INVALID_DINODE(alloc_inode->i_sb, fe);
status = -EIO;
goto bail;
}
BUG_ON((count + start_bit) > ocfs2_bits_per_group(cl));
mlog(0, "%llu: freeing %u bits from group %llu, starting at %u\n",
(unsigned long long)OCFS2_I(alloc_inode)->ip_blkno, count,
(unsigned long long)bg_blkno, start_bit);
status = ocfs2_read_block(osb, bg_blkno, &group_bh, OCFS2_BH_CACHED,
alloc_inode);
if (status < 0) {
mlog_errno(status);
goto bail;
}
group = (struct ocfs2_group_desc *) group_bh->b_data;
status = ocfs2_check_group_descriptor(alloc_inode->i_sb, fe, group);
if (status) {
mlog_errno(status);
goto bail;
}
BUG_ON((count + start_bit) > le16_to_cpu(group->bg_bits));
status = ocfs2_block_group_clear_bits(handle, alloc_inode,
group, group_bh,
start_bit, count);
if (status < 0) {
mlog_errno(status);
goto bail;
}
status = ocfs2_journal_access(handle, alloc_inode, alloc_bh,
OCFS2_JOURNAL_ACCESS_WRITE);
if (status < 0) {
mlog_errno(status);
goto bail;
}
le32_add_cpu(&cl->cl_recs[le16_to_cpu(group->bg_chain)].c_free,
count);
tmp_used = le32_to_cpu(fe->id1.bitmap1.i_used);
fe->id1.bitmap1.i_used = cpu_to_le32(tmp_used - count);
status = ocfs2_journal_dirty(handle, alloc_bh);
if (status < 0) {
mlog_errno(status);
goto bail;
}
bail:
if (group_bh)
brelse(group_bh);
mlog_exit(status);
return status;
}
/* will give out up to bits_wanted contiguous bits. */
static int ocfs2_claim_suballoc_bits(struct ocfs2_super *osb,
struct ocfs2_alloc_context *ac,
handle_t *handle,
u32 bits_wanted,
u32 min_bits,
u16 *bit_off,
unsigned int *num_bits,
u64 *bg_blkno)
{
int status;
u16 victim, i;
u16 bits_left = 0;
u64 hint_blkno = ac->ac_last_group;
struct ocfs2_chain_list *cl;
struct ocfs2_dinode *fe;
mlog_entry_void();
BUG_ON(ac->ac_bits_given >= ac->ac_bits_wanted);
BUG_ON(bits_wanted > (ac->ac_bits_wanted - ac->ac_bits_given));
BUG_ON(!ac->ac_bh);
fe = (struct ocfs2_dinode *) ac->ac_bh->b_data;
if (!OCFS2_IS_VALID_DINODE(fe)) {
OCFS2_RO_ON_INVALID_DINODE(osb->sb, fe);
status = -EIO;
goto bail;
}
if (le32_to_cpu(fe->id1.bitmap1.i_used) >=
le32_to_cpu(fe->id1.bitmap1.i_total)) {
ocfs2_error(osb->sb, "Chain allocator dinode %llu has %u used "
"bits but only %u total.",
(unsigned long long)le64_to_cpu(fe->i_blkno),
le32_to_cpu(fe->id1.bitmap1.i_used),
le32_to_cpu(fe->id1.bitmap1.i_total));
status = -EIO;
goto bail;
}
if (hint_blkno) {
/* Attempt to short-circuit the usual search mechanism
* by jumping straight to the most recently used
* allocation group. This helps us mantain some
* contiguousness across allocations. */
status = ocfs2_search_one_group(ac, handle, bits_wanted,
min_bits, bit_off, num_bits,
hint_blkno, &bits_left);
if (!status) {
/* Be careful to update *bg_blkno here as the
* caller is expecting it to be filled in, and
* ocfs2_search_one_group() won't do that for
* us. */
*bg_blkno = hint_blkno;
goto set_hint;
}
if (status < 0 && status != -ENOSPC) {
mlog_errno(status);
goto bail;
}
}
cl = (struct ocfs2_chain_list *) &fe->id2.i_chain;
victim = ocfs2_find_victim_chain(cl);
ac->ac_chain = victim;
ac->ac_allow_chain_relink = 1;
status = ocfs2_search_chain(ac, handle, bits_wanted, min_bits, bit_off,
num_bits, bg_blkno, &bits_left);
if (!status)
goto set_hint;
if (status < 0 && status != -ENOSPC) {
mlog_errno(status);
goto bail;
}
mlog(0, "Search of victim chain %u came up with nothing, "
"trying all chains now.\n", victim);
/* If we didn't pick a good victim, then just default to
* searching each chain in order. Don't allow chain relinking
* because we only calculate enough journal credits for one
* relink per alloc. */
ac->ac_allow_chain_relink = 0;
for (i = 0; i < le16_to_cpu(cl->cl_next_free_rec); i ++) {
if (i == victim)
continue;
if (!cl->cl_recs[i].c_free)
continue;
ac->ac_chain = i;
status = ocfs2_search_chain(ac, handle, bits_wanted, min_bits,
bit_off, num_bits, bg_blkno,
&bits_left);
if (!status)
break;
if (status < 0 && status != -ENOSPC) {
mlog_errno(status);
goto bail;
}
}
set_hint:
if (status != -ENOSPC) {
/* If the next search of this group is not likely to
* yield a suitable extent, then we reset the last
* group hint so as to not waste a disk read */
if (bits_left < min_bits)
ac->ac_last_group = 0;
else
ac->ac_last_group = *bg_blkno;
}
bail:
mlog_exit(status);
return status;
}
void ocfs2_populate_inode(struct inode *inode, struct ocfs2_dinode *fe,
int create_ino)
{
struct super_block *sb;
struct ocfs2_super *osb;
int use_plocks = 1;
sb = inode->i_sb;
osb = OCFS2_SB(sb);
if ((osb->s_mount_opt & OCFS2_MOUNT_LOCALFLOCKS) ||
ocfs2_mount_local(osb) || !ocfs2_stack_supports_plocks())
use_plocks = 0;
/*
* These have all been checked by ocfs2_read_inode_block() or set
* by ocfs2_mknod_locked(), so a failure is a code bug.
*/
BUG_ON(!OCFS2_IS_VALID_DINODE(fe)); /* This means that read_inode
cannot create a superblock
inode today. change if
that is needed. */
BUG_ON(!(fe->i_flags & cpu_to_le32(OCFS2_VALID_FL)));
BUG_ON(le32_to_cpu(fe->i_fs_generation) != osb->fs_generation);
OCFS2_I(inode)->ip_clusters = le32_to_cpu(fe->i_clusters);
OCFS2_I(inode)->ip_attr = le32_to_cpu(fe->i_attr);
OCFS2_I(inode)->ip_dyn_features = le16_to_cpu(fe->i_dyn_features);
inode->i_version = 1;
inode->i_generation = le32_to_cpu(fe->i_generation);
inode->i_rdev = huge_decode_dev(le64_to_cpu(fe->id1.dev1.i_rdev));
inode->i_mode = le16_to_cpu(fe->i_mode);
inode->i_uid = le32_to_cpu(fe->i_uid);
inode->i_gid = le32_to_cpu(fe->i_gid);
/* Fast symlinks will have i_size but no allocated clusters. */
if (S_ISLNK(inode->i_mode) && !fe->i_clusters)
inode->i_blocks = 0;
else
inode->i_blocks = ocfs2_inode_sector_count(inode);
inode->i_mapping->a_ops = &ocfs2_aops;
inode->i_atime.tv_sec = le64_to_cpu(fe->i_atime);
inode->i_atime.tv_nsec = le32_to_cpu(fe->i_atime_nsec);
inode->i_mtime.tv_sec = le64_to_cpu(fe->i_mtime);
inode->i_mtime.tv_nsec = le32_to_cpu(fe->i_mtime_nsec);
inode->i_ctime.tv_sec = le64_to_cpu(fe->i_ctime);
inode->i_ctime.tv_nsec = le32_to_cpu(fe->i_ctime_nsec);
if (OCFS2_I(inode)->ip_blkno != le64_to_cpu(fe->i_blkno))
mlog(ML_ERROR,
"ip_blkno %llu != i_blkno %llu!\n",
(unsigned long long)OCFS2_I(inode)->ip_blkno,
(unsigned long long)le64_to_cpu(fe->i_blkno));
set_nlink(inode, ocfs2_read_links_count(fe));
trace_ocfs2_populate_inode(OCFS2_I(inode)->ip_blkno,
le32_to_cpu(fe->i_flags));
if (fe->i_flags & cpu_to_le32(OCFS2_SYSTEM_FL)) {
OCFS2_I(inode)->ip_flags |= OCFS2_INODE_SYSTEM_FILE;
inode->i_flags |= S_NOQUOTA;
}
if (fe->i_flags & cpu_to_le32(OCFS2_LOCAL_ALLOC_FL)) {
OCFS2_I(inode)->ip_flags |= OCFS2_INODE_BITMAP;
} else if (fe->i_flags & cpu_to_le32(OCFS2_BITMAP_FL)) {
OCFS2_I(inode)->ip_flags |= OCFS2_INODE_BITMAP;
} else if (fe->i_flags & cpu_to_le32(OCFS2_QUOTA_FL)) {
inode->i_flags |= S_NOQUOTA;
} else if (fe->i_flags & cpu_to_le32(OCFS2_SUPER_BLOCK_FL)) {
/* we can't actually hit this as read_inode can't
* handle superblocks today ;-) */
BUG();
}
switch (inode->i_mode & S_IFMT) {
case S_IFREG:
if (use_plocks)
inode->i_fop = &ocfs2_fops;
else
inode->i_fop = &ocfs2_fops_no_plocks;
inode->i_op = &ocfs2_file_iops;
i_size_write(inode, le64_to_cpu(fe->i_size));
break;
case S_IFDIR:
inode->i_op = &ocfs2_dir_iops;
if (use_plocks)
inode->i_fop = &ocfs2_dops;
else
inode->i_fop = &ocfs2_dops_no_plocks;
i_size_write(inode, le64_to_cpu(fe->i_size));
OCFS2_I(inode)->ip_dir_lock_gen = 1;
break;
case S_IFLNK:
if (ocfs2_inode_is_fast_symlink(inode))
inode->i_op = &ocfs2_fast_symlink_inode_operations;
else
inode->i_op = &ocfs2_symlink_inode_operations;
i_size_write(inode, le64_to_cpu(fe->i_size));
break;
default:
inode->i_op = &ocfs2_special_file_iops;
init_special_inode(inode, inode->i_mode,
inode->i_rdev);
break;
}
if (create_ino) {
inode->i_ino = ino_from_blkno(inode->i_sb,
le64_to_cpu(fe->i_blkno));
/*
* If we ever want to create system files from kernel,
* the generation argument to
* ocfs2_inode_lock_res_init() will have to change.
*/
BUG_ON(le32_to_cpu(fe->i_flags) & OCFS2_SYSTEM_FL);
ocfs2_inode_lock_res_init(&OCFS2_I(inode)->ip_inode_lockres,
OCFS2_LOCK_TYPE_META, 0, inode);
ocfs2_inode_lock_res_init(&OCFS2_I(inode)->ip_open_lockres,
OCFS2_LOCK_TYPE_OPEN, 0, inode);
}
ocfs2_inode_lock_res_init(&OCFS2_I(inode)->ip_rw_lockres,
OCFS2_LOCK_TYPE_RW, inode->i_generation,
inode);
ocfs2_set_inode_flags(inode);
OCFS2_I(inode)->ip_last_used_slot = 0;
OCFS2_I(inode)->ip_last_used_group = 0;
if (S_ISDIR(inode->i_mode))
ocfs2_resv_set_type(&OCFS2_I(inode)->ip_la_data_resv,
OCFS2_RESV_FLAG_DIR);
}
/*
* Extend the filesystem to the new number of clusters specified. This entry
* point is only used to extend the current filesystem to the end of the last
* existing group.
*/
int ocfs2_group_extend(struct inode * inode, int new_clusters)
{
int ret;
handle_t *handle;
struct buffer_head *main_bm_bh = NULL;
struct buffer_head *group_bh = NULL;
struct inode *main_bm_inode = NULL;
struct ocfs2_dinode *fe = NULL;
struct ocfs2_group_desc *group = NULL;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
u16 cl_bpc;
u32 first_new_cluster;
u64 lgd_blkno;
if (ocfs2_is_hard_readonly(osb) || ocfs2_is_soft_readonly(osb))
return -EROFS;
if (new_clusters < 0)
return -EINVAL;
else if (new_clusters == 0)
return 0;
main_bm_inode = ocfs2_get_system_file_inode(osb,
GLOBAL_BITMAP_SYSTEM_INODE,
OCFS2_INVALID_SLOT);
if (!main_bm_inode) {
ret = -EINVAL;
mlog_errno(ret);
goto out;
}
mutex_lock(&main_bm_inode->i_mutex);
ret = ocfs2_inode_lock(main_bm_inode, &main_bm_bh, 1);
if (ret < 0) {
mlog_errno(ret);
goto out_mutex;
}
fe = (struct ocfs2_dinode *)main_bm_bh->b_data;
/* main_bm_bh is validated by inode read inside ocfs2_inode_lock(),
* so any corruption is a code bug. */
BUG_ON(!OCFS2_IS_VALID_DINODE(fe));
if (le16_to_cpu(fe->id2.i_chain.cl_cpg) !=
ocfs2_group_bitmap_size(osb->sb, 0,
osb->s_feature_incompat) * 8) {
mlog(ML_ERROR, "The disk is too old and small. "
"Force to do offline resize.");
ret = -EINVAL;
goto out_unlock;
}
first_new_cluster = le32_to_cpu(fe->i_clusters);
lgd_blkno = ocfs2_which_cluster_group(main_bm_inode,
first_new_cluster - 1);
ret = ocfs2_read_group_descriptor(main_bm_inode, fe, lgd_blkno,
&group_bh);
if (ret < 0) {
mlog_errno(ret);
goto out_unlock;
}
group = (struct ocfs2_group_desc *)group_bh->b_data;
cl_bpc = le16_to_cpu(fe->id2.i_chain.cl_bpc);
if (le16_to_cpu(group->bg_bits) / cl_bpc + new_clusters >
le16_to_cpu(fe->id2.i_chain.cl_cpg)) {
ret = -EINVAL;
goto out_unlock;
}
trace_ocfs2_group_extend(
(unsigned long long)le64_to_cpu(group->bg_blkno), new_clusters);
handle = ocfs2_start_trans(osb, OCFS2_GROUP_EXTEND_CREDITS);
if (IS_ERR(handle)) {
mlog_errno(PTR_ERR(handle));
ret = -EINVAL;
goto out_unlock;
}
/* update the last group descriptor and inode. */
ret = ocfs2_update_last_group_and_inode(handle, main_bm_inode,
main_bm_bh, group_bh,
first_new_cluster,
new_clusters);
if (ret) {
mlog_errno(ret);
goto out_commit;
}
ocfs2_update_super_and_backups(main_bm_inode, new_clusters);
out_commit:
ocfs2_commit_trans(osb, handle);
out_unlock:
brelse(group_bh);
brelse(main_bm_bh);
ocfs2_inode_unlock(main_bm_inode, 1);
out_mutex:
mutex_unlock(&main_bm_inode->i_mutex);
iput(main_bm_inode);
out:
return ret;
}
static int ocfs2_truncate_file(struct inode *inode,
struct buffer_head *di_bh,
u64 new_i_size)
{
int status = 0;
struct ocfs2_dinode *fe = NULL;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
struct ocfs2_truncate_context *tc = NULL;
mlog_entry("(inode = %llu, new_i_size = %llu\n",
(unsigned long long)OCFS2_I(inode)->ip_blkno,
(unsigned long long)new_i_size);
unmap_mapping_range(inode->i_mapping, new_i_size + PAGE_SIZE - 1, 0, 1);
truncate_inode_pages(inode->i_mapping, new_i_size);
fe = (struct ocfs2_dinode *) di_bh->b_data;
if (!OCFS2_IS_VALID_DINODE(fe)) {
OCFS2_RO_ON_INVALID_DINODE(inode->i_sb, fe);
status = -EIO;
goto bail;
}
mlog_bug_on_msg(le64_to_cpu(fe->i_size) != i_size_read(inode),
"Inode %llu, inode i_size = %lld != di "
"i_size = %llu, i_flags = 0x%x\n",
(unsigned long long)OCFS2_I(inode)->ip_blkno,
i_size_read(inode),
(unsigned long long)le64_to_cpu(fe->i_size),
le32_to_cpu(fe->i_flags));
if (new_i_size > le64_to_cpu(fe->i_size)) {
mlog(0, "asked to truncate file with size (%llu) to size (%llu)!\n",
(unsigned long long)le64_to_cpu(fe->i_size),
(unsigned long long)new_i_size);
status = -EINVAL;
mlog_errno(status);
goto bail;
}
mlog(0, "inode %llu, i_size = %llu, new_i_size = %llu\n",
(unsigned long long)le64_to_cpu(fe->i_blkno),
(unsigned long long)le64_to_cpu(fe->i_size),
(unsigned long long)new_i_size);
/* lets handle the simple truncate cases before doing any more
* cluster locking. */
if (new_i_size == le64_to_cpu(fe->i_size))
goto bail;
/* This forces other nodes to sync and drop their pages. Do
* this even if we have a truncate without allocation change -
* ocfs2 cluster sizes can be much greater than page size, so
* we have to truncate them anyway. */
status = ocfs2_data_lock(inode, 1);
if (status < 0) {
mlog_errno(status);
goto bail;
}
/* alright, we're going to need to do a full blown alloc size
* change. Orphan the inode so that recovery can complete the
* truncate if necessary. This does the task of marking
* i_size. */
status = ocfs2_orphan_for_truncate(osb, inode, di_bh, new_i_size);
if (status < 0) {
mlog_errno(status);
goto bail_unlock_data;
}
status = ocfs2_prepare_truncate(osb, inode, di_bh, &tc);
if (status < 0) {
mlog_errno(status);
goto bail_unlock_data;
}
status = ocfs2_commit_truncate(osb, inode, di_bh, tc);
if (status < 0) {
mlog_errno(status);
goto bail_unlock_data;
}
/* TODO: orphan dir cleanup here. */
bail_unlock_data:
ocfs2_data_unlock(inode, 1);
bail:
mlog_exit(status);
return status;
}
